Self-check
If you fixed one process end to end, which one should it be?
Three minutes, six questions about each of three processes, and a rough number at the end. This is the quick version of the first stage of an engagement. It estimates. It does not measure.
Your answers are never sent anywhere. They are held in the page and in the part of the link after the # symbol, which browsers do not transmit to any server. If you copy the link yourself, that copy carries your answers, which is what makes it shareable.
Pick your three most painful processes
0 of 3 chosen
The ones that eat time, get redone, or cause arguments.
Six questions about each
Wage plus everything on top. A rough figure is fine.
Currency guessed from your browser settings. Change it if that is wrong.
Illustrative example. These numbers are invented, not measured, and not from any client.
Result
Your three, plotted
Rough annual cost of leaving it alone
Rough, for prioritization, not a business case. This is what the process costs you in time, from your own inputs. It is not a saving, and nothing here claims it would be recovered.
Show the arithmeticformula and assumptions, editable
Two scores out of 100. Value is hours per week scaled to 40 points, plus rework up to 24, plus revenue upside up to 16, plus cost of an error up to 20. Feasibility is where the data lives up to 40 plus whether one person owns it up to 30, then scaled to 100. The pick is 0.75 times value plus 0.25 times feasibility. Ties go to the more feasible one, because a faster first win is worth more than a marginally larger one.
Feasibility is weighted a quarter on purpose. It matters, because a process with no records and no owner cannot be measured in two or three weeks. But it describes how a process is run today, which is not the same as how much it would repay attention, so it is allowed to decide between close calls and not to crown a process that is merely tidy.
The cost figure is hours per week times your hourly rate times the number of working weeks below, shown as a band from 0.7 to 1.3 times that. The band is a deliberate blur, not a confidence interval: there is no error propagation behind it, only a refusal to print one confident number from six rough answers.
Change either one and the numbers above update. The weighting is my judgement, not a measured model. The paid version measures instead of estimating.
Want the real version of this? Book 30 minutes.
The tool gives you the rough answer free. The paid version measures instead of estimates, and the numbers are yours either way.
What this tool does and does not do
It ranks three processes you nominate, using your own rough answers about time, rework, revenue upside, cost of failure, where the data lives, and who owns the work. Those six are a useful rough cut, not most of the decision: dependencies, regulation, seasonality, and who is already drowning matter too, and none of those fit in six questions.
It does not measure anything. Every number it shows is derived from what you typed, which means it inherits whatever is wrong with your estimate. It cannot tell you a change will pay for itself, and it never claims a saving. The figure it shows is what a process costs in time today, not what you would get back.
It can also tell you no. If the strongest of your three still scores weakly, it says so rather than crowning a weak winner, because recommending the least bad option is how these tools usually go wrong.